Skip to main content

Quick Start

1. Install and serve​

npm install -g @graphmemory/server
cd /path/to/your-project
graphmemory serve

That's it. No config file needed — the current directory becomes your project. Graph Memory will:

  1. Download the embedding model on first run (~560 MB, cached for future use)
  2. Index your project in three sequential phases: docs → files → code (models load lazily, keeping memory usage low)
  3. Start the server on http://localhost:3000

2. Open the Web UI​

Navigate to http://localhost:3000 in your browser. You'll see the dashboard with stats about your indexed project.

3. Connect an AI assistant​

Claude Code​

claude mcp add --transport http --scope project graph-memory http://localhost:3000/mcp/your-project

The project ID is your directory name (e.g., my-app for /path/to/my-app).

Claude Desktop​

Go to Settings > Connectors and add the URL:

http://localhost:3000/mcp/your-project

Cursor / Windsurf​

Add to your .mcp.json:

{
  "mcpServers": {
    "graph-memory": {
      "type": "http",
      "url": "http://localhost:3000/mcp/your-project"
    }
  }
}

4. Start asking questions​

Once connected, your AI assistant has access to 70 tools. Try asking:

  • "What does this project do?" — uses docs_search and docs_list_files to find docs
  • "Show me all exported functions in auth.ts" — uses code_get_file_symbols
  • "Create a note about the database schema" — uses notes_create
  • "What tasks are in progress?" — uses tasks_list

What happens during indexing?​

Graph Memory creates six interconnected graphs from your project:

GraphWhat it indexesWhat you can do
DocsMarkdown files → heading chunksSearch docs, browse topics
CodeTS/JS files → AST symbolsSearch code, get symbols
FilesAll files → metadataBrowse files, search by path
KnowledgeYour notes and factsCreate, search, link notes
TasksYour tasksKanban workflow, track work
SkillsReusable recipesStore and recall procedures

Next steps​